Model is not exported after running model_main_tf2.py and exporter_main_v2.py. Don't know why?

Question

I am doing object detection using TensorFlow Object Detection API in Google colab. This is my directory structure.

object_detection/
  training/
  exported_model/
  pipeline.config
  model_main_tf2.py
  exporter_main_v2.py

I run bellow for training.

!python model_main_tf2.py --model_dir=training --pipeline_config_path=pipeline.config

I run bellow for exporting model.

!python exporter_main_v2.py
--input_type image_tensor
--pipeline_config_path pipeline.config
--trained_checkpoint_dir training/
--output_directory exported_model

None of the above will produce any error but after running both I am not able to see exported model in desired directory in my case(exported_model). I didn't understand what is wrong ?

Answer 1

I found that while I run training even though it didn't produce any error It also not successful. Because It didn't generate files which should be generated after successful training like checkpoints. The training/ directory was blank.

this file's documentation helped me to figure out my problem. You can check it for further guidance.

Answer 2

i follow the struction using export_tflite_graph_tf2.py